Skip to content

Molding and Casting

Task

Group assignment:

  • Review the safety data sheets for each of your molding and casting materials
  • Make and compare test casts with each of them
  • Compare printing vs milling molds

Group assignment can be found here. I didn't contribute this week.

Individual assignment:

  • Design a mold around the process you’ll be using, produce it with a smooth surface finish, and use it to cast parts.

Ready to pour!

What could go wrong??

What could go wrong??

Lavapiés mold

Since I made the original models for Horizontes Cercanos, the Instituto Geogáfico Nacional has released updated models with more resolution. I also can use Deep Learning to increase that resolution from 2m to 0.5m, and then I could make a very high-quality print. That could be then 3D printed in the SLA printer and used as a positive for mold making.

Heightfield

I use open data from the Instituto Geográfico Nacional.

Thick barrio

I generally make the barrios shallower, but I added a bit of depth for mold making. I also made it partially hollow and punched some holes for drainage:

Punching holes

A great thing about this kind of maps is that, by definition, there can't be any overhangs. By the way the are constructed, they even have a bit of a draft angle by default!

Draft angle in heightfields

Tiny Town

I want to test the feasibility of milling moulds for my cityscapes in a CNC for increased quality compared to 3D printing, but milling a whole city would probably take too long, so I had given up.

Then I had an idea: select only a few blocks in the barrio and mill molds out of them! It could be a nice city building mosaic game.

Making the 3D model

It's important, when you export from QGIS (right click on layer / export / save as...), to select "rendered image" rather than raw data in the save dialog, so that image programs that are not specific to GIS can handle it.

I tried using the heightfield command in Rhino with the 3 different options. I felt that "surface with control points at sample locations" conserved more of the feeling of the real thing, specially at sharp angles, for example where facades meet streets.

My first iteration had too little margin on the sides so that the blocks would have had little or no adjacent streets. I went back and edited the nodes with the node tool in QGIS to fix them. That included some of the opposing blocks and also the trees so I had to fix that in the result. I tried using Krita but the results were bad. In the end, I used Rhino's align to select the control points in the sidewalk area and align them to a plane.

After careful editing of the surfaces (it was a lot more work than I expected!!) I got to a point where there were decent sidewalks. I Boolean-union'ed them into a mold of molds:

Making the mold

I milled that in the SRM-20. I did one roughing pass with a flat 1/8th inch endmill, then a finishing pass with the 1/16th flat endmill, then a final one with the 1/16th inch ball-nose endmill.

Pouring molds

I wanted to cast both edible and non-edible city blocks. I started with the food safe silicone in order not to taint the wax mold. Carefully mixed it, vacuumed it, poured it...

And it came out great! Out to the oven for degassing.

Next up was non-food-safe silicone. I picked up the closest one, a beautiful crystal clear silicone:

And after waiting for it to cure I got my awful surprise! This silicone is in no way fit for using as mold material. It breaks down way too easily so its almost impossible to demold:

After that, I poured a mold with easyplat silicone and that came out great.

Pouring pieces

Then I had the pieces ready! I first tried acrystal. It has a finish similar to plaster but it sets really fast, in about 10 minutes. It sets so fast that it actually took me by surprise the first time! I spent too much time mixing and vacuuming it. By the time I noticed it had started to become more viscous to the moment it actually set only around 30 seconds passed.

I also tried epoxi resin, both with liquid pigments and with colored micas. The results were lovely.

I especially liked playing with mixing both. For example you can set the micas on the bottom of the mold only instead of mixing them thoroughly, and it will result in a "golden snow" effect, for example:

Gotcha

Don't use the liquid pigments for epoxi on the acrystal! They make it clot almost instantly.

Stamp

I want to make 2 stamps:

  • One with my personal logo
  • One with "horizontescercanos.com" in the typography that I use for Horizontes Cercanos.
    • It would be nice to have also the main models' names so that I can stamp them in the cardboard rather than transfer them with acetone.

I've talked to Julia and she recommended different materials:

  • Laser-cut Linoleum -> It's probably great since it's used in engraving
  • Laser-cut EVA rubber (?) -> Gerard says it doesn't work well for stamping: it is too variable in its ink load
  • CNC-milled wax then silicone(s)
  • Photopolymer -> I bought some from Aliexpress

Julia recommmended Versacraft for ink pad.

Laser engraving linoleum

I bought some regular engraving linoleum from servei estació. Tested different laser cutter settings on it:

From the Rayjet 400 operating manual (pdf): "A higher DPI value improves the engraving quality but results in increased engraving depth and longer engraving time using otherwise identical parameters as the laser has to process accordingly more lines."

Reading the manual, I found that it has a dedicated stamp mode! This reflects the image (doh! I wasn't doing this!) and inverts the color (ugh). It also lets you set a shoulder angle, which supposedly helps a lot with the durability of the stamps.

To set different dpi and to acces the alternative modes like "stamp", you need to click "print usem system dialog", then Rayjet manager's "preferences" button, which will take you to the engraver driver (section 2 of the manual).

With 1000PPI, the process is really slow but the quality is seriously amazing.

Examining them closely with lenses I noticed that:

  • High speeds (80% of max) resulted in uneven depth of the cuts: the center of the path taken by the laser head goes deeper. I'm assuming that the software overcorrects for the laser heads' braking acceleration.
  • Other parameters did not result in obvious effect as long as they were changed coordinatedly: for example, power 15 speed 10 1x pass would result in similar (slightly larger) depth as power 15 speed 20 2x passes. The quality was in general not visibly affected.
  • The exception to that is that using stamp mode at 1000ppi did get better quality than using generic raster mode at 500ppi. It's also convenient for the mirroring and the shoulder.

TODOs / unpursued ideas

  • Silicone stamp with my logo -> milled wax negative or SLA? I think milled wax
  • Mold replica of Lavapiés with super-resolution
  • Tiny milled test case for barrios in the Roland
  • Silicone battery case/holder
  • Silicone magnet holder
  • Chocolate
  • Cylinder seal: with any of the techniques on a roller
  • Embossing seal?

References